Learning Camera-Aware Noise Models
This work addresses the fundamental challenge of accurate noise modeling for image processing and computer vision applications, offering a data-driven solution that is camera-specific.
The authors tackled the problem of modeling real-world imaging sensor noise, which is more complex than existing statistical models, by proposing a camera-aware generative noise model that learns from real-world data and outperforms previous methods both quantitatively and qualitatively.
Modeling imaging sensor noise is a fundamental problem for image processing and computer vision applications. While most previous works adopt statistical noise models, real-world noise is far more complicated and beyond what these models can describe. To tackle this issue, we propose a data-driven approach, where a generative noise model is learned from real-world noise. The proposed noise model is camera-aware, that is, different noise characteristics of different camera sensors can be learned simultaneously, and a single learned noise model can generate different noise for different camera sensors. Experimental results show that our method quantitatively and qualitatively outperforms existing statistical noise models and learning-based methods.